HINT: Healthy Influential-Noise based Training to Defend against Data Poisoning Attacks. (arXiv:2309.08549v1 [cs.LG])

While numerous defense methods have been proposed to prohibit potential
poisoning attacks from untrusted data sources, most research works only defend
against specific attacks, which leaves many avenues for an adversary to
exploit. In this work, we propose an efficient and robust training approach to
defend against data poisoning attacks based on influence functions, named
Healthy Influential-Noise based Training. Using influence functions, we craft
healthy noise that helps to harden the classification model against poisoning
attacks without significantly affecting the …

adversary attacks data data poisoning data sources defense exploit noise poisoning poisoning attacks research training untrusted work